Mobile app version of vmapp.org
Login or Join
Berumen635

: How to Interpret Histograms of Computed Gray Images? Situation: unable to interpret histogram of flat images in XnView where no labels on axes where I think y-axis = pixel number, and x-axis

@Berumen635

Posted in: #ImageEditing

Situation: unable to interpret histogram of flat images in XnView where no labels on axes where I think y-axis = pixel number, and x-axis = [0,255]
Original image: just sin(x) wave going on the gray area
Images (similar original image and histogram)




I did not find anything about the documentation to interpret it, thread here.
Those dotted peaks are about interval lengths in pixel scale [0,255].

Open


Why two thin peaks for the gray image? I know that the other peak is caused by the colorbar. The other peak must be caused by the sin wave. - - How can you get labels on the axis such that I can evaluate better it?


System: Linux Ubuntu 16.04 64 bit
Hardware: Macbook Air 2013-mid
Image editing: XnView 0.81

10.01% popularity Vote Up Vote Down


Login to follow query

More posts by @Berumen635

1 Comments

Sorted by latest first Latest Oldest Best

 

@Michele215

Edited new answer

What is an histogram?

An histogram is a graph that shows the relation between the brightness of a pixel versus the number of pixels that has that brightness. In this case, the level of gray.



Comparing the graph with it's photo, some times you can identify the peaks of the graph to some zones on your photo:



Depending on the content of the image, for example in a low key photo (a dark one) the graph piles up to the left.



In the image on the right I am forcing a lot more the darkness of the image. The brighter value is now on the red arrow, and you can see how the graph piles up more and more. You still can see some parts of the image (yellow arrow).

On the case of flat images with biiiig areas of one level you will see only a (proportionally) very tall thin line on your graph, because you have no transitions, no shades, just pure contrast.

Almost in all programs, (and cameras that has the feature) the graph scales up and down relative to the highest value. If you have a lot of varity of shades, you will see clear curves as in the photo one.

But when you have a very dominant tone, as the second photo, the rest of the values get flatter, proportionally of the increasing height of the tallest peak (meaning a looooooot of pixels has that value).

Practical aplications

In photography and design it has a lot of aplications, from seeing the overall tone and "mood" of the image, to some very specific warnings.

In photography you normally never want to "clip" the bright side, overexposing the highlights. It is an irreparable mistake.

In graphic design for print you normally want to avoid an image that has no black, because will look dull when printed. Depending on the image, a white point can be necessary too, so you extend that histogram till the darker point and lighter point touches a bit the sides of the graph.



Old answer

I tested my version of XnView on windows and I have no problem viewing a normal histogram.

If the image you are viewing is a flat black square you will have an histogram that consist on a single tall line staked to the left... that you probably won't see.

Try a normal b&w photo.

10% popularity Vote Up Vote Down


Back to top | Use Dark Theme